Shafin Haque

Hi! I'm Shafin, a 19 y/o at UC Berkeley studying Electrical Engineering and Computer Sciences (EECS). I'm interested in machine learning, computer vision, and computational photography.

This past summer I interned at Glass Imaging where I developed deep learning models for smartphone image signal processors to enhance mobile camera quality.

Check out my work and experience on this page, and feel free to reach out!

Email  /  LinkedIn  /  Github  /  Devpost

We propose Box Prediction Rebalancing for single-stage object detectors to combat the challenge of learning from partially annotated datasets. By randomly removing percentages of negative predictions from our model's loss computation, our model performs better as it learns less from false positives which may be true species of interest without ground truth.

• Deep learning for mobile image signal processors, working on image restoration tasks such as denoising, demosaicing, enhancement, and super-resolution.
• Lead face restoration project, training efficient SOTA image restoration models with techniques specifically for faces such as GAN-inversion with high quality facial priors.
• Worked directly with RAW sensor captures, enhancing real details while balancing hallucinations.

• Developed few-shot model for semantic segmentation of workpiece objects in automotive manufacturing lines. Trained model with only 15 images per camera site.
• Built light-weight single-shot object detector using features from existing production post estimation model for face blurring.

• Novel object detection methods for partially labeled datasets, streamlining identification and counting process of marine species in underwater ROV videos.
• Wrote extended abstract accepted to NeurIPS LMLR and contributed to research paper accepted to International Journal of Computer Vision and CVPR Workshop.

More of my projects can be found on my Github or Devpost.

• Python web app for predicting burn area, put-out time, and location of wildfires with machine learning. Used CNNs, regression, and object detection detection models
• 4th place Congressional App Challenge 2021 & 4x Hackaton Award Winner

• RShiny app for predicting the next pitch in baseball based on current game scenario. Used regression and neural network based approaches
• Wrote article published in TowardsAI documenting process

Website template from here